National Baseline on Civil Registration and Vital Statistics in Cambodia

National Baseline on Civil Registration and Vital Statistics in Cambodia Implemented by the General Department of Identification / Ministry of Interior October 2016-November 2016 Supported by UNICEF Technical support by Oliver Schell

Table of Contents List of Tables... II List of Figures... II Abbreviations... III Statistical fact sheet, figures at a glance... IV Acknowledgements... V Executive Summary... 1 1. Background... 5 2. Overall Objective of the Baseline Survey... 6 2.1. Specific Objectives... 6 3. Limitations... 6 4. Methodology... 7 4.1. Sample Selection... 7 4.2. Set up of Survey Team... 8 4.3. Development of Baseline Survey Tool... 9 4.4. Data Management and Statistical Analysis... 9 5. Findings...10 5.1. Sample Size and Profile...10 5.2. Quality of Maintaining CRVS Records in Communes...15 5.3. Birth Registration...16 5.3.1. Overall Results...16 5.3.2. Regional Differences...17 5.3.3. Gender...19 5.3.4. Distance to Commune Offices...20 5.3.5. Ethnicity...21 5.3.6. Residential Status...22 5.3.7. Parent s Education...23 5.3.8. Health Service Provision...24 5.3.9. Economic Status...25 5.3.10. Disability...25 5.3.11. Identification Card Issuing Campaigns...25 5.4. Marriage Certificate...26 5.5. Death Certificate...28 6. Conclusions...30 7. Recommendations...31 Attachments...32 I

List of Tables Table 1: Sample distribution among the five target provinces of the baseline survey...10 Table 2: Rural and urban population sample...11 Table 3: Disability including combination of different impairments...14 Table 4: Comparing birth registration among age groups post mobile registration campaign and within mobile registration campaign....17 Table 5: Birth registration rates for children under five and children under one year...17 Table 6: Differences in birth registration between rural and urban areas...18 Table 7: Overall birth registration disaggregated by gender...20 Table 8: Mean and median distance to commune...20 Table 9: Cluster size of close and far villages...20 Table 10: Comparison of birth registration between Khmer and non-khmer ethnic groups...22 Table 11: Influence of residential status on birth registration...23 Table 12: Comparison of birth registration rates between higher and lower educational levels...24 Table 13: Comparison of birth registration of children delivery by skilled health staff and TBAs....24 Table 14: Birth registration compared between ID poor and non poor families....25 Table 15: Birth registration among people with disability compared to people without disability...25 Table 16: Marriage certificates...26 Table 17: Count of under-age partnerships...27 Table 18: Marriage certificates and level of education...28 Table 19: Possession of death certificates...28 Table 20: Registered deaths and reasons for deaths per province...29 Table 21: Reasons of death registered...29 List of Figures Figure 1: Population pyramid of the total sample...10 Figure 2: Ethnic groups represented in the survey...12 Figure 3: Residential status of households...13 Figure 4: Location of delivery...13 Figure 5: Deliveries by skilled health staff compared to TBAs...14 Figure 6: Quality of maintaining civil registry twin books...15 Figure 7: Reporting of health centres to communes...16 Figure 8: Overall birth registration rate including birth certificates and certified birth...16 Figure 9: Birth registration rates by region comparing cluster of people covered by the mobile registration campaign with children born after registration campaign...18 Figure 10: Multiple risk factor analysis per province...19 Figure 11: Birth registration rates disaggregated by distance from village to commune office...21 Figure 12: Birth registration rates of different ethnic groups in comparison...22 Figure 13: Birth registration rates in relation to educational level of parents...23 Figure 14: Different birth registration rates according to location of delivery...24 Figure 15: Percentage of people in a partnership having marriage certificates per age groups...27 II

Abbreviations CRVS CCWC e.g. ESCAP GDI ID poor MOI MOP NSPI RAF TBA Civil Registration and Vital Statistics Commune Committee for Women and Children Example given Economic and Social Commission for Asia and the Pacific General Department of Identification Identification of poor households programme of the Ministry of Planning Ministry of Interior Ministry of Planning National Strategic Plan for Identification Regional Action Framework Traditional Birth Attendants III

Statistical fact sheet, figures at a glance Indicator Total sample population with registered birth (Regional Action Framework Target 1.C) Population registered birth who were covered by mobile registration campaign (2004-2006) Population registered birth who were born after the mobile registration campaign 0-9 years = post-mobile campaign sample Percent Significance (total numbers) (Chi square) 89.7% (19983/22274) Not applicable 92.0% (16539/17970) 80.2% (3414/4259) P= <0.001 Children <1 registered birth (RAF Target 1.A) 74.4% (390/524) Not applicable Children <5 registered birth (RAF Target 1.B) 78.0% (1711/2193) Not applicable Children <18 registered birth 84.6% (6444/7613) Not applicable Male of total sample population registered birth 90.0% (9926/11025) Female of total sample population registered birth 89.4% (10057/11249) P= 0.123 Birth registration in village with distance to commune within 91.7% (11696/12761) median range 0-2 km Birth registration in village with distance to commune above 87.1% (8287/9531) P= <0.001 median range <2.1 km Birth registration -in post mobile campaign sample- in village 83.0% (1969/2373) with distance to commune within median range 0-2 km Birth registration -in post mobile campaign sample- in village 76.6% (1444/1885) P= <0.001 with distance to commune above median range <2.1 km Urban population of total sample registered birth 97.7% (6895/7055) Rural population of total sample registered birth 86.0% (13088/15219) P= <0.001 Khmer of total sample registered birth 93.7% (18073/19298) Non Khmer (indigenous people and ethnic minorities) 64.2% (1910/2976) P= <0.001 Population of total sample registered birth with residential 92.6% (17180/18544) documents (family book, registration at location) Population of total sample registered birth without residential 75.2% (2776/3692) P= <0.001 documents (family book, registration at location) Children of mother from higher education registered birth * 92.5% (1495/1616) Children of mother from lower education registered birth * 72.3% (1804/2494) P= <0.001 Children of father from higher education registered birth * 91.1% (1825/2003) Children of father from lower education registered birth * 68.8% (1328/1929) P= <0.001 Children delivered by skilled health staff registered birth* 86.3% (2965/3436) Children delivered by TBA registered birth * 54.4% (442/813) P= <0.001 Children of non- poor families registered birth* 81.3% (2931/3604) Children of ID poor card holders registered birth* 72.6% (431/594) P= <0.001 People with no disability registered birth 89.8% (19680/21907) People with disability registered birth 83.5% (298/357) P= <0.001 Couples, separated or widows with marriage certificates 29.9% (3325/11124) Not applicable Deceased with death certificates (RAF Target 1.D) 46.9% (60/128) Not applicable *sample: post mobile campaign IV

Acknowledgements This national baseline on civil registration which by sample size is one of the largest baselines of its kind worldwide- is the result of a great team effort of five different departments within the General Department of Identification (GDI). More than 30 staff from the following departments participated in its implementation: The Department of Civil Registration, the Department of Administration, the Department of Population Statistics, the Department of Identification Cards, the Department of Passports and the Department of Information Systems. As technical adviser to the baseline survey I would like to express my gratitude to His Excellency Mao Chandara, General Director of GDI, His Excellency Yin Malyna, Deputy General Director of GDI and Mr. Eng Chandara, Director of the Department of Civil Registration for their strong support giving priority to the baseline implementation to all departments. I would also like to express my gratitude to the two team leaders Mr. Heng Sophat and Mr. Oeung Kim Unn for their leadership in the field and during data entry, and also to Mr. Yim Sam Ol for his senior advice regarding civil registration activities in Cambodia. Especially I would like to personally thank Mr. Kamphorn Sathya for his restless efforts to facilitate the smooth implementation of the survey at all steps and for the translation of the report into Khmer. Finally appreciation goes to UNICEF s Community Development team for their efforts to support the implementation of this baseline survey. Key persons within UNICEF were: Mr. Yi Kosalvathanak, Ms. Alexandra Graf, Mr. Tomas Jensen, Ms. Keo Sovannary, and Ms. Anne Lubell. V

Executive Summary Introduction In October and November 2016 the General Department of Identification of the Ministry of Interior conducted a baseline survey on civil registration to define absolute levels of Civil Registration and Vital Statistics (CRVS) in line with the National Strategic Plan for Identification and requirements of the Economic and Social Commission for Asia and the Pacific (ESCAP). The implementation of the survey was made possible with financial support from UNICEF. The survey provides baseline data on key indicators that now enable GDI and stakeholders to objectively monitor CRVS progress, and to plan targeted and specific interventions to improve civil registration. To identify gaps and factors that contribute to inequality in registration rates, the data was disaggregated by relevant indicators such as: gender, geography, ethnicity, residential status, ID poor status, educational status, location of birth and disability. The method of data collection was to gather information on birth, marriage and death from individuals from randomly selected households of the general population. The survey coverd in four provinces and the capital Phnom Penh. Birth registration The baseline survey captured data on 22416 individuals from 4681 households. There are significant differences of registration rates of the people who were covered by the Cambodian mobile registration campaign conducted in 2004-2006, when more than 90 per cent of the population were registered 1, and children who were born after the mobile campaign who today are at the age of 0-9. The overall birth registration rate including those who were covered by the mobile registration campaign is 89.7 per cent (19983/22274; 142 missing data on birth registration). The registration rate for children below the age of 18 is 84.6 per cent (6444/7613). The registration rate of children who were born after the mobile registration campaign (age of 0-9), which represents the current performance of the CRVS system in Cambodia, is 80.2 per cent (3414/4259). For children under the age of five the rate is 78.0 per cent (1711/2193) which is 4.7 per cent higher than the results of the Cambodian Demographic Health Survey 2014, and the registration rate of children born within the past 12 months is 74.4 per cent (390/524). There are significant differences in birth registration rates in urban and rural areas and among groups who have certain vulnerability factors. The survey has shown that geographic areas or social groups 1 Ministry of Interior, Report of Birth Registration Results after Mobile Registation Campaign (Khmer). 1

where vulnerability factors accumulate are more likely to show low registration rates. Identifying those areas and groups will help to plan individualized interventions. Vulnerability factors: Living in a rural area Urban population of total sample registered birth 97.7% (6895/7055) Rural population of total sample registered birth 86.0% (13088/15219) P= <0.001 Living far from the commune office Birth registration in village with distance to commune within median range 0-2 km Birth registration in village with distance to commune above median range <2.1 km Belonging to indigenous groups or ethnic minorities 91.7% (11696/12761) 87.1% (8287/9531) P= <0.001 Khmer of total sample registered birth 93.7% (18073/19298) Non Khmer (indigenous people and ethnic minorities) 64.2% (1910/2976) P= <0.001 Families who do not have residential status at the location the baby is born Population with residential documents (family book, registration at location) Population without residential documents (family book, registration at location) 92.6% (17180/18544) 75.2% (2776/3692) P= <0.001 Children born into families with lower educational level Father from higher education 91.1% (1825/2003)* Father from lower education 68.8% (1328/1929)* P= <0.001 *sample: post mobile campaign Children not delivered by skilled health staff Children delivered by skilled health staff registered birth 86.3% (2965/3436)* Children delivered by traditional birth attendant 54.4% (442/813)* P= <0.001 *sample: post mobile campaign Families who are ID poor holders 2

Children of non- poor families registered birth* 81.3% (2931/3604)* Children of ID poor card holders registered birth* 72.6% (431/594)* P= <0.001 *sample: post mobile campaign People with disability People with no disability registered birth 89.8% (19680/21907) People with disability registered birth 83.5% (298/357) P= <0.001 Marriage certificates Among 11124 individuals living in relationships, had separated or were widowed only 29.9 per cent (3325/11124) had a marriage certificate. It is widely acknowledged that people still lack understanding of the benefits of a marriage certificate and rather perceives it as an obstacle in case the partnership splits and a divorce has to be filed. The awareness about relevance of marriage certificates seems to increase with people s education which is the most important correlation factor. Among people with no education only 12.5 per cent (236/1889) had a marriage certificate, while the highest rate of marriage certificates of 76.5 per cent was among individuals who had gone to university. Death certificates Certification of death is still less than half, 46.9per cent (60/128) of cases, but higher than marriage registration. The rate might be higher because the certificate is to be used to e.g. prove property claims such as land titles, bank accounts etc. or to present it to employers when employees asked for absence from work to attend a funeral. In most cases, 89.4 per cent (371/415) a reason of death was registered. However the reasons registered were in many cases not appropriate to be used for health statistics. In 145 cases the reason of death was simply stated as disease, in 78 cases as old age which sums up to 63.2 per cent of all reasons of death entered. None of the reasons had been accompanied by a medical certificate. Recommendations Set lower individual regional targets for areas like Ratanakiri where there are many factors that have shown a significant effect on reduced birth registration Increase early birth registration (within 30 days) e.g. by awareness raising campaigns that also reach populations with lower education: Mass media such as radio or television, peer education or messages to mobile phones and smart devices. 3

Strengthening the role of the CCWC and are the cooperation with health care providers are an opportunity to link delivery to timely birth registration. Conduct mobile campaigns in certain areas until 2024 to compensate for barriers to birth registration. Targeting should focus on specific groups such as indigenous people, ethnic minorities, areas with many ID poor households, urban poor communities, migrant workers etc. Link campaigns or services to the possession of civil registration certificates. Two examples are: school enrolment and issuing an identity card. For both a precondition is to have a birth certificate or a certified birth certificate. This means that people use and need them, it increases their importance because a practical use and benefit are connected. Continue to ensure user friendly and equitable access to CRVS documents for all. 4

1. Background In September 2015 the United Nation s General Assembly issued the new Agenda for Sustainable Development. Sustainable Development Goal 16.09 covers civil registration: By 2030, provide legal identity for all, including birth registration. A regional commitment to foster civil registration was made by the Economic and Social Commission for Asia and the Pacific (ESCAP) at a Ministerial Conference on Civil Registration and Vital Statistics (CRVS) in Asia and the Pacific, which was held in November 2014 in Bangkok. Ministers proclaimed The Asian and Pacific CRVS Decade (2015-2024) and a Ministerial Declaration to Get Every One in the Picture was adopted. The Ministers endorsed and committed to the implementation of a Regional Action Framework (RAF) for civil registration. The conference defined three main goals for the action framework: Goal 1: Universal civil registration of births, deaths and other vital events; Goal 2: All individuals are provided with legal documentation of civil registration of births death and other vital events, as necessary, in order to claim identity, civil status and ensuing rights Goal 3: Accurate, complete and timely vital statistics (including on causes of death), based on registration records, are produced and disseminated Civil registration in Cambodia started in the 1920 th during the French colonial rule, it continued during the reign of Norodom Sihanouk as head of state and the Lon Nol regime until 1975, however during these periods civil registration was widely limited to urban areas. Cambodia s registration records were completely destroyed under the Pol Pot regime and until 2002 there was no standardized legislation on civil registration. The registration process in Cambodia started in 2002, but until 2004, only 300,000 people were registered or less than five per cent of the total population. In October 2004, the Ministry of Interior launched a Nation-Wide Mobile Civil Registration Campaign. The mobile registration resulted in over 90 per cent nation-wide birth registration rate, representing 11 million people who had their births registered, by the end of the campaign in December 2006. 2 The MOI subsequently succeeded to maintain a high registration of vital events and to improve their services. Between 2010 and 2014 Cambodian Demographic Health Surveys show a significant increase in the registration of children under five from 62.1 per cent (N=8122) in 2010 to 73.3per cent (N=7805) 3 in 2014 respectively. 2 UNICEF,LGCR, Concept Note- Working towards Improving Registration of Vital Events in Cambodia. 3 Statistics/Cambodia, Health/Cambodia, and International, Cambodia Demographic and Health Survey 2014. 5

However, by 2016 the General Department of Identification (GDI) of the Ministry of Interior responsible for CRVS did not yet possess any reliable baseline data on the general birth registration rate including adults or on marriage or death registration. The National Strategic Plan for Identification (NSPI) commits that: based on the existing data and in line with the ESCAP requirements and methodology the General Department of Identification will work on defining a CRVS coverage baseline that will facilitate measuring the progress towards established CRVS coverage goals. During the first phase GDI will define absolute levels of CRVS coverage. 4 Living up to this commitment of the NSPI is the aim of our baseline survey. 2. Overall Objective of the Baseline Survey The overall objective of the baseline survey is to define absolute levels of CRVS coverage in line with the requirements of the National Strategic Plan for Identification and the ESCAP Regional Action Framework on CRVS. The survey will provide baseline data on key indicators that will enable MOI and stakeholders to objectively monitor progress, and to plan targeted and specific interventions to fill gaps of civil registration. 2.1. Specific Objectives To provide quantitative data on registration rates of birth, marriage and death To disaggregate registration rates by relevant indicators to analyse potential gaps and necessary fields of interventions. 3. Limitations The purpose of the survey according to NSPI is to define absolute levels of CRVS coverage and is thus designed as quantitative survey. Although the survey has identified gaps and vulnerable groups through disaggregating data by key indicators, and also provides recommendations that result from the interpretation of quantitative data, the author will leave it to GDI and its partners to draw additional conclusions. The report focussed on the general population and thus followed a strictly random sampling methodology (despite applying some pre-defined criteria as proposed by ESCAP). Therefore the survey did not specifically target vulnerable populations to avoid sampling bias. Groups that were not directly 4 MOI, National Strategic Plan for Identification 2017-2026. 6

covered include: children in institutional care, urban poor communities or mobile communities of fisher-folk on the Tonle Sab etc. 4. Methodology 4.1. Sample Selection To gather quantitative data on CRVS the survey targeted randomly selected households of the general population. Information was collected from individuals in household interviews. The definition of a household followed that of the Ministry of Planning s Identification of Poor Household Programme as: All people who eat from the same rice pot or share money for food. A household could thus consist of blood related persons, or non blood related, even friends or other people living together. Persons had to live in the household for at least three months to be counted a household member. Target provinces were sampled randomly but clustered into four predefined criteria: metropolitan area rural areas remote areas with indigenous population and border areas The survey covered five provinces, accounting for 20 per cent of the 25 provinces/municipalities. Within each province, initially six communes were selected representing 10 per cent of an average of 60 communes per province. In each commune interviewers covered two villages accounting for 20 per cent of an average of ten villages per commune. The following main five target areas were selected Metropolitan City: Phnom Penh as the only metropolitan city in Cambodia Border provinces: Two border provinces with main border crossings were nominated for random selection: Svay Rieng and Banteay Meanchey province. Svay Rieng was selected. Remote province with high proportion of indigenous population: Four provinces were nominated: Ratanakiri, Mondulkiri, Stoeung Treng, and Preah Vihear. Ratankiri was randomly selected. Rural Areas: All remaining provinces were eligible for random selection. Odormeanchey and Preah Sihanouk were selected. The capital city of Sihanouk Ville was excluded from the random sampling for rural areas. Map of randomly samples provinces 7

Source: www.mangomap.com, WFP Humanitarian Response Forum The remaining selection of districts, communes and villages was done with a research randomizer available online (www.randomizer.org) and the Ministry of Planning Identification of Poor Household list of provinces, districts, communes and villages to assign the numbers to target areas. Taking financial and human resource limitation into account the sample size calculation followed a logic based on the number of data collectors that could be employed, how many households they could visit in a day, multiplied by a week of field work in each province. This produced a total of 4500 households targeted. 900 households per province. Because CRVS data was collected on individuals in each household the total sample size was estimated by multiplying with the average number of household members: 4500 households * 4.6 mean size of households 5 = 20700 individuals. The estimated sample size was exceeded during the survey which achieved a total size of 22416 individuals. 4.2. Set up of Survey Team A Baseline Survey Team was established according to the instructions of General Director of Identification. It was led by one Deputy General Director and two team leaders of whom one was responsible for data collection and the other one for data entry and analysis. Technical assistance was provided by a part-time consultant. Interviewers were recruited from different departments of the GDI 5 Cambodia Demographic and Health Survey 2014. 8

and trained in data collection. Data entry in the SPSS data bases was done by GDI staff and analysis was done jointly facilitated by the consultant. 4.3. Development of Baseline Survey Tool The information collection tools cover the key indicators of the ESCAP Regional Action Framework. Despite data on birth, marriage and death registration the tool covered social- and economic indicators taking the Cambodian context into consideration. The design of the questionnaire followed a consultative process with GDI core staff, UNICEF representatives and other stakeholders. Key indicators included in the questionnaire were: Age in years for age group disaggregation Sex to assess gender discrepancies Distance of village to commune office Ethnicity to assess differences among ethnic groups Local residential status to capture situation of migrants who are not registered at the location of interview Educational status to analyse impact of educational level of parents on children s birth registration Location of birth to compare birth registration between children born with assistance of skilled health staff to traditional birth attendants ID poor status to assess influence of economic status on registration rates Disability Despite the individual household interview an additional questionnaire for village chiefs was designed to capture general information such as distance and road conditions to commune. A third information collection tool was for the commune level focussing on death registration and the capacity of communes to perform CRVS activities. 4.4. Data Management and Statistical Analysis Data was entered into a SPSS data base. Frequency analyses including percentages were performed for each variable. Cross tables on dichotome variables performing Pearson chi square tests were done to assess statistical significance of variables on registration rates. We defined that a p-value of below 5 percent (p <0.05) as statistically significant. Linear regression analysis were performed where appropriate. 9

5. Findings 5.1. Sample Size and Profile The baseline survey provides data on 22416 individuals from 4681 households, which results in an average household size of 4.8 which is in line with Ministry of Planning (MOP) data from the Inter- Censal Population Survey 2013 with similar figures. 6 The total sample was evenly distributed among the five main target provinces with Preah Sihanouk having the lowest sample size. Table 1: Sample distribution among the five target provinces of the baseline survey Province Frequency Percent Ratanakiri 4391 19.6% Odormeanchey 5111 22.8% Preah Sihanouk 3578 16.0% Svay Rieng 4616 20.6% Phnom Penh 4720 21.1% Total 22416 100.0% The mean age of individuals registered in the survey is 28.1 years. The median age is 25.0 which complies with the findings of the Inter Censal Population Survey 2013 with a median age of 24.5 years. 7 The population pyramid shows a drop in population towards the 40-45 year age cluster and a postconflict baby boom with a sharp growth of the population until the beginning of the 1990s. Then the pyramid shows an expected demographic transition with a decrease in birth rates among the clusters of 15-25 years of age. In our sample there has been a recent increase in births especially during the past five years- however those trends will need to be verified by census surveys in the future. Figure 1: Population pyramid of the total sample 6 Cambodian Inter-Censal Population Survey 2013. 7 Ibid. 10

[100-105) [90-95) [80-85) [70-75) [60-65) [50-55) [40-45) [30-35) [20-25) [10-15) [0-5) -1500-1000 -500 0 500 1000 1500 Female Male N= 22362, 54 cases missing The survey captured 2202 children under five amounting for 9.8 per cent of the total sample and 520 children under one year. According to GDI estimates, 92 per cent 8 of the population had been registered as a result of the mobile campaign from 2004-2006. This excellent registration rate during the mobile campaign reflects in the overall birth registration rates of the baseline survey. In order to judge the performance of today s civil registration mechanisms we will especially look at the cluster of children who are today 0-9 years, the post mobile registration campaign sample. The post mobile campaign sample has a size of 4276 individuals or 19.1 per cent of the total sample of the survey. Comparing results of the post campaign cluster with the population that was covered by the mobile campaign is helpful because it shows that many disadvantaged populations significantly benefit from mobile registration activities. About one third of the population was defined as urban, meaning Phnom Penh municipality and three Sangkat (urban communes) in Svay Rieng town. The other two third were defined as rural. Table 2: Rural and urban population sample Frequency Valid Percent Rural 15319 68.3 Urban 7097 31.7 Total 22416 100.0 N=22,416, no missing 8 Ministry of Interior, Report of Birth Registration Results after Mobile Registation Campaign (Khmer). 11

The gender distribution of the sample is almost equal with 50.5 per cent (11314) female and 49.5 percent (11102) male individuals. To evaluate the economic status of households we asked if they were holders of the Identification of Poor Households Equity Card called the ID poor card issued by the Ministry of Planning. Within our sample population 12.7 percent (2786/21997, 419 missing) were ID poor holders. Another important variable for disaggregation of data was ethnicity. A pre -defined criterion for the selection of one target province in the survey was to have a high proportion of indigenous people. The randomly selected province of Ratanakiri has a majority indigenous population and is also home to many ethnic Laotians who traditionally make up a strong ethnic group in the Northeast of Cambodia. Another ethnic minority group that is sufficiently represented in the survey are the Cham. The Cham are descendants of the Champa kingdom. They are the Muslim community in Cambodia. Figure 2: Ethnic groups represented in the survey 1758; 582; 164; 522; 8% 3% 1% 2% Khmer Cham 19390; 86% Indigenous People (Tampouen, Jarai, Krung, Phnong, Kouy) Ethnic Minority (Lao,Vietnamese, Chinese) Mixed and other ethnicities N=22,416, no missing Vietnamese are represented with only 50 individuals, 0.2 percent of the total sample. The random selection of survey sites shows a low proportion of Vietnamese, even though two provinces Ratanakiri (9 individuals) and Svay Rieng (1 individual)- directly border Vietnam. Phnom Penh had a sample of 29 Vietnamese. Migration plays an important role in birth registration. Children have to get their birth certificate from the commune where their parents are registered as residents. If a family moves away from the commune for work, it can be a great burden for them to travel back to their home commune to get their children registered. In order to capture this dilemma we asked households if they were registered at the location of the household interviewed or if they were living here but being registered elsewhere. 12

Figure 3: Residential status of households 3737; 16,7% No residential status 18632, 83.3% Having residential status N=22369, 47 missing The province with the highest proportion of people without residential status was Ratanakiri with 33.8 per cent (1476/4365). The survey team interviewed many families who are migrant workers in the province s agricultural industries. The highest rate of residential registration was in Phnom Penh were 90.6 per cent (4267/4710) were registered. GDI seeks a strong cooperation with health service providers in order to guide parents to register their children. Therefore it is relevant to know where people were born and if there is a correlation with registration rates. Sub national local governance structures like the Commune Committees for Women and Children (CCWC) build a linkage between health services and local administrations. Reporting on deliveries in Health Centers is an agenda of the CCWC at the commune meeting. Figure 4: Location of delivery 62; 0% 14288; 65% 5339; 25% 679; 3% 1497; 7% Public health facility Private Clinic At home by skilled midwife At home by TBA Other N=21865, 551 missing The majority of people in the sample were delivered with the assistance of traditional birth attendants (TBA). Actually the proportion of deliveries by skilled staff surpassed TBA s only 10-15 years ago. 13

Figure 5: Deliveries by skilled health staff compared to TBAs 87.4% 100.0% 80.0% 60.0% 40.0% 12.6% 20.0% 0.0% Skilled health staff Traditional birth attendant N=21823, 593 missing Even today in the 0-5 age cluster still a significant proportion of 12.6 per cent are delivered by TBA despite this practice has officially been abandoned. This figure however is due to the high proportion of 40.6 per cent of the 0-5 year olds (214/527) who were still delivered by TBAs in Ratanakiri. In the other provinces these rates were much less: 3.4 per cent in Odormeanchey, 7.7 per cent in Preah Sihanouk, 3.4 per cent in Svay Rieng and 1.8 per cent in Phnom Penh. The rate of disabled people in the survey sample was 1.6 per cent (361/22416). This rate is lower than the prevalence of disability assessed in the last Cambodian Demographic and Health Survey. Table 3: Disability including combination of different impairments Type of disability Frequency Valid Percent Seeing 109 30.2 Hearing 36 10.0 Walking 113 31.3 Memory, concentration 21 5.8 Self-care 18 5.0 Contact/communication 34 9.4 Using arm/hand 4 1.1 Hearing and seeing 11 3.0 Hearing and walking 2 0.6 Seeing and walking 5 1.4 Seeing and memory concentration 1 0.3 Walking, self-care 2 0.6 More than two impairments combined 5 1.5 Total 361 100.0 14

The CDHS gives a prevalence of disability of any kind with 9.5 per cent. This includes also disabilities that result only in having only some difficulties. The CDHS category for serious disability defined as a lot of difficulty or cannot do has a prevalence of 2.1 per cent. 9 During the baseline survey we asked for level of disability that inhibits the performance of daily tasks which will more likely represent the latter CDHS category of serious disablitiy. 5.2. Quality of Maintaining CRVS Records in Communes At 21 commune offices an inspection of quality of registration activities was done by checking the registration twin books. More than half of the communes scored good or very good. All six communes in Svay Rieng scored good (5) or very good (1).The three weak scores come from Ratanakiri (1 very weak) and Oddar Meanchey (2 weak). Figure 6: Quality of maintaining civil registry twin books 12 10 8 6 4 2 0 10 7 1 2 1 very good good average weak very weak More than 80 per cent of communes receive reports on deliveries in health centers from the public health services. The common occasion of reporting is the monthly meeting at the commune level when health centre representatives attend commune meetings. Many topics are discussed in this general monthly meeting and the number of children delivered in health facilities is one agenda point reported by Commune Committee for Women and Children member. In the 5 communes who did not receive a report, it was because there is no health centre in the commune. 9 Cambodia Demographic and Health Survey 2014. 15

Figure 7: Reporting of health centres to communes 5, 19% 22, 81% Health center reported Health center no report N=27 communes 5.3. Birth Registration 5.3.1. Overall Results The overall birth registration rate among the total sample population in the baseline survey was 89.7 per cent. This figure however includes all ages and thus also those who were registered during the mobile registration campaign 2004-2006. Figure 8: Overall birth registration rate including birth certificates and certified birth 2291; 10.3% 19983; 89.7% No registration Registered birth N=22,274, 142 missing The table below disaggregates between the sample covered during the mobile registration campaign and the post mobile campaign sample which reflects on the current performance of CRVS systems. We see that the registration rate among those aged 0-9 years is 80.2 per cent. Compared to the 92.0 per cent that had been registered during the mobile campaign, the drop of about 12 per cent to the current performance is statistically significant with a Pearson Chi-square p-value of <0.001. 16

Table 4: Comparing birth registration among age groups post mobile registration campaign and within mobile registration campaign. Sample Post mobile registration 0-9 years Within mobile registration 10 years and above No birth Registered registration birth Total 845 3414 4259 19.8% 80.2% 100.0% 1431 16539 17970 8.0% 92.0% 100.0% Total 2276 19953 22229 Total per cent 10.2% 89.8% 100.0% N=22229, 187 missing = 54 missing ages and 133 unknown status of birth registration. Pearson Chi-square test on statistical significant difference p<0.001 Even though statistically significance is shown, the programmatic reality of maintaining an 80.2 per cent birth registration among the total population is very good. The registration rate for children below the age of 18 is 84.6 per cent (6444/7613). Results for the younger children however show lower registration rates with 78.0 per cent for children under five and 74.4 per cent for children that were born within the past year. A reason for this might be that many parents still register their children when the birth certificate is demanded for school enrolment at the age of six. Table 5: Birth registration rates for children under five and children under one year. Children <5 years registered birth 78.0% (1711/2193) Children <1 year registered birth 74.4% (390/524) In Cambodia there is a difference between certified birth certificates and birth certificates which are only issued within the first thirty days of life. There were only three new-borns in the survey sample below one month. Of those one already had a birth certificate. However there were 99 babies of the age of one month. Among those 67 had a birth certificate, 5 had a certified birth certificate and 27 did not yet have their birth registered amounting for an overall registration rate of 72.7 per cent at the age of one month to two months. 5.3.2. Regional Differences Disaggregating the birth registration data by province shows regional differences between the five sample provinces. In the chart below the blue bars represent the sample cluster that was covered by the mobile registration campaign, the red charts children between 0-9 who were born after the mobile registration. 17

Figure 9: Birth registration rates by region comparing cluster of people covered by the mobile registration campaign with children born after registration campaign 97.7% 96.5% 98.4% 95.5% 92.1% 78.7% 95.7% 86.8% 73.0% 52.0% 100.0% 80.0% 60.0% 40.0% 20.0% Phnom Penh Svay Rieng Preah Sihanouk Odormeanchey Ratanakiri 0.0% Post mobile registration Within mobile registration N=22229, 187 missing The municipal sample of Phnom Penh reaches close to 100% birth registration. It has managed to maintain a high registration rate even after the mobile registration campaign with a drop of only 1.2 per cent (97.7 per cent to 96.5 per cent). Svay Rieng achieved similar good results with 98.4 per cent during the mobile registration and maintains 95.5 per cent in the post mobile registration sample. A drop of only 2.9 per cent. Svay Rieng is a small province with high population density and thus a dense net of administrative centers. Impressive results also in Odormeanchey, selected as a rural area and a remote north-western province of Cambodia. Odormeanchey achieved 95.7 per cent during the mobile registration campaign and maintains 86.8 per cent, a drop of less than 10 per cent (8.9 per cent). Preah Sihanouk was sampled as rural province and achieved 92.1 per cent during the mobile registration and maintains 78.7 per cent which is a drop of 13.4 per cent. The worst result of the baseline survey has Ratanakiri where only 73.0 per cent were achieved in the mobile campaign and within the post mobile campaign sample only 52.0 per cent are maintained. In general the data shows that there is a statistically significant lower birth registration in rural areas compared to urban. Defined as urban were Phnom Penh and three Sangkats in Svay Rieng provincial capital town. Table 6: Differences in birth registration between rural and urban areas Rural Urban Total Pearson Chi-square <0.001 No Registered registration birth Total 2131 13088 15219 14.0% 86.0% 100.0% 160 6895 7055 2.3% 97.7% 100.0% 2291 19983 22274 10.3% 89.7% 100.0% 18

The baseline survey has looked at a number of vulnerability factors to analyse if they contribute to lower civil registration rates. The report reflects on the relations of: gender, distance, ethnicity residential status, education, health services, economic situation, disability and ID card issuing on civil registration activities. The results show that a combination of multiple factors exacerbates low registration rates. The clearest example is Ratanakiri Province. Ratanakiri outruns the other provinces in possessing a number of risk factors that could contribute to its low birth registration. It has a majority of non-khmer population consisting of indigenous people and ethnic Lao and the lowest educational status. Because many migrants from low-land Cambodia work on Rubber plantations and other agricultural industries, Ratanakiri also has the highest rate of people without having residential status. And finally it has the highest rate of deliveries still performed by Traditional Birth Attendants. Figure 10: Multiple risk factor analysis per province from post mobile registration campaign sample cluster 0% 20% 40% 60% 80% Ratanakiri Odormeanchey Non Khmer population No residential status Preah Sihanouk Svay Rieng Low or no education of father TBA delivery Phnom Penh N varies within each sub category The following paragraphs will assess the influence of those and other indicators on birth registration in more detail. 5.3.3. Gender Gender inequality is an issue in many countries, however Cambodia shows gender equality with regards to birth registration. There is statistically no significant difference in the registration results with 90.0 per cent of male and 89.4 per cent of females registered. The gender equality was also maintained after the mobile registration campaign with a female registration rate of 79.6 per cent (1614/2028) and male registration rate of 80.7 per cent (1800/2231). We can conclude that Cambodia 19

is a good role model for other countries to ensure that there is no discrimination between the sexes with regards to CRVS. Table 7: Overall birth registration disaggregated by gender Gender Male Female Total No Registered registration birth Total 1099 9926 11025 10.0% 90.0% 100.0% 1192 10057 11249 10.6% 89.4% 100.0% 2291 19983 22274 10.3% 89.7% 100.0% Chi square p=0.125 confirming no statistically significant difference. 5.3.4. Distance to Commune Offices A general assumption is that remoteness and distance to commune offices has an impact on birth registration. In our sample the average distance from village to commune was 3.36 km, the median distance was 2 km. The mean was increased by extremely remote villages in Ratanakiri and Odormeanchey. In Ratanakiri the median was double the national average and Odormeanchey had the farthest average distance. Table 8: Mean and median distance to commune Distance from village to commune Mean km Median km Ratanakiri 4.29 4 Odormeanchey 4.92 3 Preah Sihanouk 1.77 1 Svay Rieng 3.49 1 Phnom Penh 1.89 1 Total 3.36 2 To analyse the influence of distance on birth registration the sample was disaggregated into two groups. We looked at the birth registration rate of close villages within the median range of zero to two kilometres and villages farther than the median distance. Table 9: Cluster size of close and far villages Number of individuals Percent Close (0-2] km within median 12761 57.3% Far (>2.1 to highest] km 9513 42.7% Total 22274 20

The bar chart below shows birth registration rates disaggregated by close village and far villages within different sample clusters: the overall sample (red bars), the sample of people who have been covered by the mobile registration campaign (green bars) and the children born after the mobile registration campaign (blue bars). Figure 11: Birth registration rates disaggregated by distance from village to commune office 100.0% 90.0% 80.0% 91.7% 87.1% 93.6% 89.7% 83.0% 76.6% 70.0% 60.0% (0-2] km (11696/12761) (2.1 -highest] km (8287/9513) (0-2] km (9727/10388) (2.1-highest] km (6843/7628) (0-2] km (1969/2373) (2.1 -highest] km (1444/1885) Overall sample (p<0.001) Within mobile registration campaign (p<0.001) Post mobile registration campain (p<0.001) In all three clusters the difference in birth registration rates between close villages within median range and farther villages is statistically significant. However the rate among populations in remote villages that were covered by the mobile campaign is with 89.7 per cent 13.1 per cent better that the 76.6 per cent registration rate in remote villages after the mobile campaign. So a mobile campaign is a good compensator for distance barriers, but in general we can conclude that people living in remote villages are less likely to be registered than people living closer to a commune office. However there is Odormeanchey province with many remote villages which perform very good. The village of Sleng Por for a example has aan overall registration rate of 96.1 per cent (393/409) and maintains 96.2 per cent in the post mobile campaign cluster (76/79) is 18 km from the commune office and is the farthest of all villages in the whole baseline survey. So distance and remoteness seem not to have a stand-alone effect on registration rates but the regional differences in birth registration are created by a combination of different social factors. 5.3.5. Ethnicity There are 24 indigenous groups recognized by the Ministry of Planning in Cambodia who make up 1.3 per cent of the total population. In the northeast of Cambodia however, including the province of Ratanakiri represented in the survey, indigenous people make up the majority of the population. Besides indigenous people significant numbers of ethnic minorities like the Cham, the Lao and Vietnamese have lived in Cambodia for many generations/centuries. Looking at birth registration rates 21

among the ethnic Khmer and the Cham population in comparison to other ethnic groups we see significant differences. Figure 12: Birth registration rates of different ethnic groups in comparison 100.0% 93.7% 92.9% 80.0% 60.0% 57.0% 59.0% 40.0% 20.0% 0.0% Khmer N=19298 Cham N=522 Indigenous Ethnic Minorities People N=1730 N=563 The cross table below shows differences between Khmer and non-khmer populations. Non Khmer include the Cham, indigenous people, ethnic minorities and also children from mixed marriages. Table 10: Comparison of birth registration between Khmer and non-khmer ethnic groups No Registered registration birth Total Khmer 1225 18073 19298 6.3% 93.7% 100.0% Non Khmer 1066 1910 2976 or mixed 35.8% 64.2% 100.0% Total 2291 19983 22274 10.3% 89.7% 100.0% Pearson Chi -square <0.001 5.3.6. Residential Status In order to register their children at a commune or sangkat office, parents need to be registered residents there. It means they have to possess a residential card or a family book that is issued in the respective area of administration. Thus migrants who do not have changed their residential status have no opportunity to register their children. This poses a significant burden to families who cannot afford the money and time to travel back to their home communes to register their children. The result of the survey shows that residential status of a family has a significant influence on birth registration. Among people who did not possess residential documents the registration rate was 75.2 per cent, whereas the registration among the people with residential documents was 92.6 per cent. Svay Rieng was selected as a border province with a lot of migration, however problems of civil 22

registration seem not to arise in the provinces that people leave, but in those areas to where people move to. Table 11: Influence of residential status on birth registration No residential status Having residential status Total Pearson Chi square p=<0.001 No Registered registration birth Total 916 2776 3692 24.8% 75.2% 100.0% 1364 17180 18544 7.4% 92.6% 100.0% 2280 19956 22236 10.3% 89.7% 100.0% 5.3.7. Parent s Education A common assumption is that awareness about importance of civil registration and the motivation to register their children is related to the educational level of parents. Thus the survey asked for educational level of fathers and mothers and indeed there is a strong correlation of increased birth registration with the increase of educational level of parents. Figure 13: Birth registration rates in relation to educational level of parents 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 58% 58% 73% 79% 88% 91% No Education Primary School Secondary School 93% 94% 97% 98% R² = 0.9179 High School University Fathers Mothers Linear (Fathers) N varies in each category, sample from post mobile registration campaign, pre-school not displayed because small sample linear increase R²=0.92 To test statistical significance we created two clusters. One cluster are children from parents with low educational level which includes: no education, pre-school, primary school and skill training and a second cluster of higher education which includes: secondary school, high school and university. The result shows a statistically higher birth registration among families with higher education. 23